Development of CRUD thickness estimation model under pressurized water reactor condition

Abstract

As interests in high burn up are increasing, Axial Offset Anomaly (AOA) receives attentions and is considered as one of the key issues that we are facing. Therefore, it is necessary to clearly understand the thermal hydraulic performance of CRUD, which is one of root causes of AOA, as it affects heat and mass transfer mechanisms on cladding surface. The KAIST AOA experiment has been conducted under Pressurized Water Reactor (PWR) operating condition. It successfully generated CRUD deposited on the test rod surface of the heated section. Also, post-examinations have been performed to measure thickness and porosity, and to analyze the composition of the CRUD. In this study, sensitivity studies were done for CRUD characteristics, and it was concluded that CRUD characteristics, except CRUD thickness, are not taking a significant role in regarding the overall heat transfer coefficient. The three-CRUD regime model is suggested to develop an empirical correlation of the overall heat transfer coefficient for CRUD thickness estimation using WALT experimental database: early CRUD regime, transition CRUD regime, and fully developed CRUD regime. In the early cruded regime, the overall heat transfer coefficient has a constant value which is the same as that estimated with the clean cladding, since the CRUD is not thick enough to interrupt the heat transfer on cladding surface. On the other hand, when CRUD thickness exceeds a critical thickness, the overall heat transfer coefficient decreases as CRUD gets thicker. In the fully developed cruded regime, the overall heat transfer coefficient has a constant value as CRUD thickness increases beyond a threshold value. The KAIST CRUD thickness estimation model is also validated with the data of the KAIST AOA experiment. Through the comparison with experimental data of the crud thickness from the KAIST AOA experiment and the WALT experiment, it turns out that the model can predict them within 30.96% of average Root Mean Square Error (RMSE).